Machine Learning Techniques for Behavioral Health Detection from Social Media Data

There is a rising prevalence of behavioral health challenges like depression, anxiety, suicidal ideation, mental disorder etc. among young people globally. Individual's apathy to attend regular clinical assessments due to cost, busy schedules, fear of stigma, technical limitations of traditional diagnostic methods, prevents early discovery of behavioral health assessment. This significant detection gap has been explored by researchers seeking more effective ways to swiftly detect, control, and avert dangers associated with indefinable behavioral health challenges. With the prevalence of social media platforms and their popular use for multipurpose information sharing, it has been identified as an unobtrusive data source, containing rich, real-time linguistic and behavioral markers of individual's psychological state, capable of providing information with potential of revolutionizing early intervention. Consequently, machine learning and natural language processing (NLP) techniques have been deployed to create predictive models for early detection of behavioral health challenges from social media data. This comparison highlights some works in this direction with a focus on their proposed techniques.